Portal de Periódicos da CAPES

Fortune telling in biliary atresia: What is in the tea leaves?

2009; Lippincott Williams & Wilkins; Volume: 15; Issue: 8 Linguagem: Inglês

10.1002/lt.21740

1527-6473

Peter F. Whitington,

Gallbladder and Bile Duct Disorders

Predicting, prognosticating, and fortune telling—they are one and the same. Biliary atresia (BA) consistently brings misfortune to the lives of children unfortunate enough to have it. Before the invention of the portoenterostomy procedure by Dr. Kasai in the 1960s, the misfortune was indeed great—100% mortality in early childhood—and fortune telling was, therefore, extremely easy. Dr. Kasai's invention substantially reduced the misfortune due to BA: a child with BA could expect an extension of life outside of early childhood with about 50% probability and even hope to reach adulthood, although the probability of that was low. The great advance occurring in the intervening 4 decades that has changed the fortune of a child with BA is the emergence and evolution of orthotopic liver transplantation (OLT). Current 5- and 10-year survival rates for OLT recipients with BA exceed 90%, no matter what the age of the child is at the time of transplant. OLT is, therefore, a terrific, life-saving fallback to portoenterostomy in the management of BA, and together they provide extended survival beyond early childhood in >95% of cases. In addition, fortune telling is again easy, at least in terms of predictably living beyond early childhood. Despite these therapeutic successes, those involved in the care of children with BA still feel the need to tell fortunes. What if the parents of a child newly diagnosed with BA want a prediction of whether the child will need a transplant and when? The “whether” is pretty easy: any baby with BA has a greater than 80% probability of needing a liver transplant given today's data and no major therapeutic breakthrough. The “when” is not so easy: only about half of patients with BA requiring transplant will need it before 2 years of age. In other words, predicting at the time of diagnosis whether a child with BA will require a transplant before 2 years is the same as predicting “heads” on a single coin toss. In this issue of Liver Transplantation, Pape et al.1 read the “tea leaves” represented by scars in liver biopsies obtained at the time of diagnosis in 53 children with BA to determine if their pattern improved the prediction of outcome in these circumstances. BA, biliary atresia; OLT, orthotopic liver transplantation. Liver biopsy interpretation is closely akin to reading tea leaves: it is highly subjective and very dependent on experience. In order to improve liver biopsy interpretation in this setting, the authors focused on one aspect of the biopsy, namely fibrosis, that they thought might be related to outcome. They used picrosirius red staining in combination with computerized morphometry to interrogate liver biopsies obtained at the time of diagnosis to provide objectivity. As a result of this analysis, each biopsy was assigned a score (Vfib) that represented the relative amount of fibrosis that each contained. These biopsies were acquired over a 5-year period (1998-2003), but they underwent picrosirius red staining all at one time for the purposes of this study. Presumably, the management of the cases was not influenced by the degree of fibrosis observed at the time, and thus the relationship between Vfib and the primary outcome (survival without transplantation) was considered to be unbiased. It became necessary to translate the continuous variable Vfib into a binary one—less fibrosis/more fibrosis—with the receiver operating characteristic approach to determine its relationship to prognosis, which could be only good or bad. The main finding of the study was that a liver at the time of diagnosis with a Vfib score of 2.5% predicted poor outcome (death or transplant) in less than 1 in 5 cases. In other words, the results of the study show that if a baby with BA shows little fibrosis on the initial biopsy, he can be expected to have a good outcome from the Kasai operation, whereas a baby with more than a little fibrosis may have a poor outcome but most likely will not. It is fair to ask who would benefit from improved fortune telling and to what degree. In other words, what was the unmet need addressed in this study? The authors suggest that improved prognostication would have value to the individual patient in that it might help guide the “individual therapeutic strategy.” It is hard to see how the study was designed with that objective in mind. Restructuring the therapeutic strategy would be important only to babies who were not going to do well with the standard one. Thus, it would seem more reasonable to have done the receiver operating characteristic analysis with the idea of accurately predicting those with poor outcome and to not worry about predicting good outcome. The authors allude to benefit to the system in that improved prognostication could improve the allocation of resources, such as donor organs, and the timing of transplantation. That is just not the case. All decent organs will be allocated because they are inadequate in number to serve the need. Because predicting outcome does not change it, more than 80% of children with BA can be expected to need a transplant at some point, so the overall burden of disease on the system does not change. Finally, the timing of OLT is determined by the status of the patient and the availability of the graft. Unless improved prognostication can lead to an accurate prediction of the status of the patient in time so that a reservation can be made for a graft, like making a hotel reservation on this or that day, and graft availability can be somehow scheduled like hotel rooms, it makes no sense to attempt fortune telling for this purpose. Although the authors met with a certain degree of success in terms of objectifying the reading of tea leaves, the results do not provide the “crystal ball” needed for fortune telling with the accuracy required to apply it in the clinical setting. It seems reasonable to expect that the extent of scarring will be related to outcome in any progressive liver disease and that BA will be no exception. BA can be an especially energetic disease when it comes to causing fibrosis. Bezerra's group recently presented, at the 2008 Annual Meeting of the American Association for the Study of Liver Diseases, findings that suggest gene-programmed differences in infants with fibrosing BA and those with more inflammatory (less fibrosing) disease.2 The fibrosing molecular subtype comprised 27 of 48 subjects and was strongly associated with death or OLT before 2 years of age (odds ratio = 8.2, 95% confidence interval = 0.85-424). The authors conclude that finding this molecular signature may be “predictive” of outcome. Unfortunately, the predictive value of the test barely reached significance (P = 0.049) and comes nowhere close to being of clinical value. It is interesting to note that a semiquantitative scale was used to judge fibrosis in Bezerra et al.'s study and that such a scale (the Ishak score3) did not correlate with Vfib in the study by Pape et al.1 Thus, it would be of considerable interest to apply quantitative picrosirius red staining to tissues examined for the molecular signature of fibrosis to determine the agreement between the approaches. Furthermore, putting the findings of the 2 studies together leads to the interesting question of why some patients with BA with extensive liver fibrosis and/or the molecular makeup for fibrotic liver disease escape the misfortunate outcome of death or OLT before the age of 2 years. Could something be discovered in the tea leaves of the liver biopsies and RNA that would add to or subtract from the finding of fibrosing disease to signal a fortune worth telling? Time will tell with good fortune.